Schema for scholarship

[mikehquan19]

Designing the schema for new datasource (scholarships, study abroad programs, etc.) should again be the collective effort and responsibility of the team rather than some individual works that don't really yield any fruits.

We will get started with designing the schema for Scholarship. Note that we rely on the 2 different sources for our scrapers:

• Scholarship listings: shows detailed information about each scholarship it lists
• Atermis: The new scholarship management system by UTD which, despite being more of a place for students to submit application and manage their scholarships, does list some of opportunities not shown in the UTD listings.

The problem with Atermis is that it lists the scholarship's info in a less structured, more natural language-based format that will be very challenging to parse unless we need some help from LLM. Nontheless, I believe that if we reverse-engineer the site, we will perhaps find some API or datasource for us to work with.

Throw your ideas here about the schema, then I will finalize it once we have the schema that everyone agrees on.

[simar-rekhi]

Adding on to what @mikehquan19 mentioned... I found some inconsistency with the listings too. Scholarship listing tool had some general scholarships (disbursed by more than two schools per session), however, this was not available on that school's personal scholarship website. As an example, I was using the Raytheon Scholarship that goes for several majors across diff. schools. So, it might be a good idea to scrape individual school offerings to:

• simply append to the main schema
• validate the content of scholarship listings

The only thing I am mostly concerned about in this scenario is if there is a scholarship visible on the school portal but not on the Listing's tool. In that cases it might be a good idea to leverage the scraped data for both the use cases jotted above

[mikehquan19]

We will have to leverage as many sources of possible. This will complicate the scraper but it is worth it (well, to certain extent). Can you provide us links to some school's personal scholarship websites? That way everyone doesn't have to look around for the source later.

[Yogeb-tech]

Since each site seems to have inconsistencies, could we just have separate scrapers (like one for Scholarship listings, another for Artemis) that all output to the same schema, with some fields being optional depending on the source? It feels a lot more scalable to separate logic for each site in case there are changes/issues with a specific website or we want to add new sources.

[mikehquan19]

@Yogeb-tech , exactly, like the events scraper like we have, which is splitted into Mazevo, Astra, Calendar.

[simar-rekhi]

here are all the scholarship listings on each school's website

• NSM: https://nsm.utdallas.edu/scholarships/
• Harry Bass: https://bass.utdallas.edu/scholarships/
• BBS: https://bbs.utdallas.edu/academics/scholarships-awards/
• EPPS: https://epps.utdallas.edu/future-students/scholarships/
• ECS: https://engineering.utdallas.edu/academics/undergraduate-majors/scholarships/
• Interdisciplinary: https://is.utdallas.edu/about-us/scholarships/
• JSOM: https://jindal.utdallas.edu/student-resources/scholarships/

[KBui4]

ECS also has a separate website for graduate scholarships.

• ECS Graduate: https://engineering.utdallas.edu/academics/graduate-education/graduate-funding-opportunities/

[mikehquan19]

What I'm thinking right now is we use the UTD listing as the base for designing the base schema for scholarships first. And then when we look into the Atermis and the schools' personal websites, we will modify the schemas continuously to accomodate with all the sources at the end. What do yall think?

[KBui4]

If we're basing the main scholarship schema off of the UTD listings, then for fields, we can have name, school, academic program (or majors), status (open/close), and deadline from the table. From the More Information page, we can also get the term, matriculation status (continuing or incoming student), citizenship status, Texas residency status, course load (full time or part time), student classification (undergrad or grad student), concentrations, and contact info. We can probably also get the additional eligibility information/requirements as well.

[mikehquan19]

Can you provide us with json schema?

[KBui4]

@greeshiee has the JSON schema with the information I mentioned.

[greeshiee]

Scholarship = {
"id": number,
"name": string,
"description": string,
"deadline": string,
"term": string,
"matriculation_status": string,
"citizenship": string,
"texas_residency_status": string,
"course_load": string,
"eligible_classification": string,
"schools": string,
"majors": string,
"concentrations": string,
"contact": string,
"application_instructions": string,
"application_link": string
}

[simar-rekhi]

we might wanna add a schema object that tells you the exact address of the scraping. Something like...
"scrape_dest" : { scholarship_listing, artemis, indiv_website }

as an additional modality, further explain indiv_website as NSM, ECS, Harry Bass, etc.

[SoggyRhino]

I think we should nest the requirements, also some of the requirements might be better represented as arrays. Definitely schools and majors but maybe some of the others

Scholarship = {
"id": number,
"name": string,
"description": string,
"deadline": string,
"term": string,
"contact": string,
"application_instructions": string,
"application_link": string,
"requirements" : {
     "matriculation_status": string,
     "citizenship": string,
     "texas_residency_status": string,
     "course_load": string,
     "eligible_classification": string,
     "schools": string[],
     "majors": string[],
     "concentrations": string[],
     }
}

[LuisAlar]

I agree with the nesting.
I was thinking about using Enums for fields like citizenship to make filtering easier. As well as update the wording to "Immigration_status", is it can allow us to include F-1, H-4 and other more specific criteria for this requirement.

The scholarships I've been seeing tend to have renewals as well, some more specific than others. Adding a field to indicate if it can be renewed could be helpful.

Would the "eligible_classification" be used to store the min GPA for the scholarship? Adding a field for it if not seems necessary.

Scholarship = {
"id": number,
"name": string,
"description": string,
"deadline": string,
"term": string,
"contact": string,
"application_instructions": string,
"application_link": string,
"requirements" : {
     "matriculation_status": string,
     "immigration_status": enum,
     "texas_residency_status": string,
     "course_load": string,
     "eligible_classification": string,
     "min_gpa": string // or double ?
     "schools": string[],
     "majors": string[],
     "concentrations": string[],
     }
 "renewal": string  
}

[SoggyRhino]

Looking into Atermis, it appears that we can get a good bit without llm parsing. Using the filter system we can get the requirements for each scholarship.

(filters are a little scuffed, some have an explicit not considered value while others just use absence)

Need-Based Award: Student must have medium or high unmet financial need through the FAFSA, TASFA or UTD Financial Statement to be considered for the scholarship/fellowship
Student Type: Continuing graduate student
                         | Continuing undergraduate student
                         | Incoming undergraduate first time in college
                         | Incoming undergraduate transfer student
                         | New to UTD graduate student or returning UTD student who is a new graduate student
Academic Career: Graduate Student
                         | Undergraduate student
Academic Department: No specific academic department required for scholarship/fellowship
                         | Scholarship/fellowship either requires or prefers applicants be pursuing a degree in the School of Economic, Political and Policy Sciences
                         | Student's academic program is within the Naveen Jindal School of Management
                         | Student's program of study must be part of the Erik Jonsson School of Engineering and Computer Science
                         | Student's program of study must be part of the Harry W Bass Jr School of Arts Humanities & Technology
Citizenship: Citizenship not a criteria
                         | Recipient can be an international student
                         | US Citizen
                         | US Permanent Resident
State Residency: Texas residency not required
                         | Texas residency required
Academic Program: Scholarship/Fellowship requires student to be in a STEM program
Disability: Scholarship/fellowship may consider a student with a mental disability
                         | Scholarship/fellowship may consider a student with a physical disability
Family Dynamics: Student is a single parent
                         | Student was an orphan prior to turning 18
                         | Student was previously in Foster care
                         | Student will be the first in their immediate family to earn an undergraduate degree
GPA Requirement: Either a minimum GPA is required, or GPA is considered

There will need to be some clean up but it looks pretty similar to Scholarship listings.

How to get it:

• Filter .category_row
  • Id ("123456")
  • Key ("Citizenship")
  • Value ("Citizenship not a criteria")
• Scholarship .result
  • Hidden markt tag contains a list of ids

[KBui4]

I was looking through Artemis and checking to see what the general UTD scholarship website does not have. This is an incomplete list, but here's what I found so far.

Missing from general UTD scholarship site:
Axxess Scholarship for the Jonsson School

• Listed in Artemis
• Listed in ECS scholarship site

BankUnited iCARE Scholarship Fund

• Listed in Artemis

BMEN IAC

• Listed in Artemis

Dean's Dissertation Completion Award

• Listed in Artemis
• Listed in Harry Bass scholarship site

Dean's MA Scholarship

• Listed in Artemis
• Listed in Harry Bass scholarship site

Dr. Gopal Gupta Undergraduate Scholarship for Computer Science established by ACM 2020

• Listed in Artemis
• Listed in ECS scholarship site

Dr. Roger N. & Adrienne T. Rosenberg Graduate Fellowship

• Listed in Artemis
• Listed in Harry Bass scholarship site

Emma and Vittorio Argento MS'76 PhD '89 Scholarship

• Listed in Artemis
• Listed in ECS scholarship site

England Family Scholarship in Engineering and Computer Science

• Listed in Artemis
• Listed in ECS scholarship site

Excellence in Education Doctoral Fellowship

• Listed in Artemis
• Listed in ECS scholarship site

Gil and Frances Lease Endowed Scholarship

• Listed in Artemis
• Listed in ECS scholarship site

Goldman Sachs Excellence in Computer Science Scholarship

• Listed in Artemis
• Listed in ECS scholarship site

HWBJ Graduate Research Award

• Listed in Artemis
• Listed in Harry Bass scholarship site

Duplicate in Artemis (maybe?)
Greater Texas Foundation - Tech Titans Removing Educational Barriers Endowment
Greater Texas Foundation Removing Educational Barriers Scholarship

It seems that if a scholarship is missing from the general UTD scholarship site, it's likely to be from ECS or Harry Bass.

Additionally, I noticed that the general site also merges all of the undergraduate and graduate JSOM scholarships into one (Naveen Jindal School of Management Undergraduate Scholarships and Naveen Jindal School of Management Graduate Scholarships).

[KBui4]

Looking into it a bit more, there's some more cases where multiple scholarships are grouped into one. For example, the Sustainable Earth Systems Sciences Scholarship listing consists of the following:

• Dallas Gem & Mineral Society Endowed Scholarship Fund
• Victor Oppenheim Scholarship Fund
• Dr. and Mrs. John G. Scales Endowed Scholarship Fund in Geosciences
• James B. Urban Scholarship
• Dr. James L. Carter Master’s Scholarship
• Dr. Matthew Brzostowski Scholarship
• Anton Hales Endowed Fellowship in Geophysics
• Dr. Oscar Wilhelm Memorial Student Research Fund
• Kristian Soegaard Memorial Fund
• Bob and Margie Rutford Fellowship